Understanding the complexity of knowledge integration in collaborative new product development teams: A case study
Post on 05-Sep-2016
Embed Size (px)
Understanding the complexity of knowledge integration incollaborative new product development teams: A case study
Maaike Kleinsmann *, Jan Buijs, Rianne Valkenburg
Department of Product Innovation Management, Delft University of Technology, Faculty of Industrial Design Engineering,Landbergstraat 15, 2628 CE Delft, The Netherlands
Consider a telephone. From the 1940s till the 1990s, a phone did not changemuchwith regard to itsproduct functionality and product architecture (Valkenburg, 2000). However, in the last 15 years, theproduct had changed in many ways. The switch from a shared house phone to a personal phone hashad a signicant impact on both the product appearance aswell as product use. People do not use theirpersonal phone only for calling, but also for taking pictures, making movies, playing games, listeningto music and as an agenda. This has changed the consumer attitudes towards a phone. What once wasa rather functional product has become a product that expresses ones personality. Therefore,customers required their phone to be modern, easy to use, and of good quality. In the meantime, theydid not want the phone to be expensive, as they would like to buy a new one every 2 years.
These customer needs reect increasing customer demands and increasing product complexity,requiring a far-reaching integration of the different knowledge domains of the actors from different
J. Eng. Technol. Manage. 27 (2010) 2032
A R T I C L E I N F O
Available online 8 April 2010
Collaborative new product development
A B S T R A C T
Knowledge integration is important in collaborative new product
development (Co-NPD). The research literature shows that the way
actors create a shared understanding about the new to create
products is a quality indicator of Co-NPD. This study investigates
what factors inuence the creation of a shared understanding in Co-
NPD. The results show factors at three different levels; the actor,
project and company level. Additionally, there exist relationships
across the factors. The related factors form four different types of
interfaces. The interfaces differ from each other since different
types of collaborative mechanisms exist within them.
2010 Elsevier B.V. All rights reserved.
* Corresponding author. Tel.: +31 15 2788657.
E-mail address: firstname.lastname@example.org (M. Kleinsmann).
Contents lists available at ScienceDirect
Journal of Engineering andTechnology Management
journal homepage: www.elsevier.com/locate/jengtecman
0923-4748/$ see front matter 2010 Elsevier B.V. All rights reserved.doi:10.1016/j.jengtecman.2010.03.003
disciplines during collaborative new product development (Co-NPD) (Garcia and Calantone, 2002;Swink, 2000; Veryzer, 2005). Granstrand et al. (1992) found that the rst generation of cellular phonesin the 1980s only required electronic skills to be developed. However, in the 1990s, the developmentof the third generation of these phones demanded knowledge in physical ergonomics, electronic,mechanical, and computer engineering, and physiological and psychological aspects. In the last 10years, the knowledge needed to develop todays mobile phones had increased yet again.
Since the increased product complexity requires knowledge and skills from different elds, Co-NPD teams within companies have the responsibility to integrate these different knowledge bases(Edmondson and Nembhard, 2009). Knowledge integration appeared to be difcult for them,because actors often had different, interests and perspectives on the new to develop product (Bondand Ricci, 1992; Bucciarelli, 1996; Dougherty, 1992; Emmanuelides, 1993; Grifn and Hauser,1996; Moenaert and Souder, 1990; Moenaert et al., 2000). Furthermore, it was demanding forcompanies manage the actors towards thinking along the same line since actors faced difcultiesin interpreting and understanding each others knowledge (Adams et al., 1998; Bucciarelli, 2002;Dougherty, 1992; Grifn and Hauser, 1996; Hage et al., 2008; Moenaert and Souder, 1990). Thereason for this was that actors from different disciplines used different languages and differentrepresentations of the design artefact. Additionally, they often had different interests andresponsibilities. Therefore, negotiations and trade-offs were required to make the actors effortscoherent (Bucciarelli, 1996).
In the case study presented in this paper, Co-NPD is dened as: the process in which actors fromdifferent disciplines share their knowledge about both the development process and its content. Theydo that in order to create shared understanding on both aspects, to be able to integrate and exploretheir knowledge, and to achieve the larger common objective: the new product to be designed(Kleinsmann, 2006). Knowledge sharing is a critical intervening variable in organizationalperformance in Co-NPD (Mohrman et al., 2003), which shows the importance of an effective Co-NPD process. Research showed that companies face some major knowledge management challengesduring Co-NPD since it thrives upon the exploration of knowledge under a high degree of uncertainty(Berends et al., 2007). Since actors do not have to become experts in each others elds, they do nothave to share all their knowledge. Yet, they have to be able to integrate their knowledge bases in asensible manner.
Therefore, Berends et al. (2007) claimed that, during Co-NPD, knowledge management shouldfocus on knowledge integration instead of focusing on knowledge transfer. They also statedthat the effectiveness of knowledge integration required a mutual understanding of theactors contributions. This is in line with research on design communication, which nds thatthe quality of the Co-NPD project is dependent on the process of creating a shared understanding(Dong, 2005; Kahn, 1996; Valkenburg, 2000). In addition, our interviews with product managersrevealed that managers need more knowledge on how they could manage the creation of a sharedunderstanding among actors from different disciplines, as the following quote from a productmanager illustrates:
. . .There is no problem that cannot be solved technically. At least I have never experienced thatpersonally. Finally, they always solve it. They use projectmanagement tools tomap the problemand to nd the solution. However, the real problem is the communication among the teams. Didthey understand each other and are they critical enough to each other? . . . Dependent on theirdiscipline, they have a feeling for each others discipline, but the real details . . . It is complex tobe able to know all parts about every discipline. As a Project Manager, you really have to beaware of this. That is also the reasonwhy I need something tomonitor that . . . (ProjectManagerPhilips).
The study presented in this paper focuses on the process of creating a shared understanding in Co-NPD projects. The paper starts with a literature review about the role of shared understanding in Co-NPD projects. It continues with a description of the case study executed, followed by a description ofthe learning history method that was used as the research method. The paper concludes with itsresults, limitations, implications for future research and with a discussion of the results andimplications for managerial practice.
M. Kleinsmann et al. / Journal of Engineering and Technology Management 27 (2010) 2032 21
2. The role of shared understanding in Co-NPD projects
In this paper the Co-NPD process was considered as a social process. According to Weick andRoberts (1993), an actor executes three main activities during this social process. The rst activity isthe construction of the task that an actor needs to perform. This allows the actor to understand that thetask is a part of a system and that others perform different and complementary tasks in the system.Thus, an actor interrelates his action to the actions of others. This shows that the relations among theactors are not predened, but constructed during the process and that actors depend on ndings fromthe other actors. A result derived by an actor performing a development task may result in thenecessity that another actor needs to make an iterative loop. Cross (1990) found that a Co-NPD teamcannot prevent these loops, since they are the nature of developing. However, these iterative loops canalso originate from mistakes in the extraction and re-enactment sense to the system. These mistakesmay indicate that a shared understanding about the design content does not exist and that the tuningacross the actors fails (Mohammed and Dumville, 2001; Mohrman et al., 2003; Valkenburg, 2000).
Mohammed and Dumville (2001) claimed that actors have created a shared understanding if theyhave reached cognitive consensus, which refers to the similarity among the actors about how the keyissues in the development process are conceptualized. Both Mohammed and Dumville (2001) andMohrman et al. (2003) added that process related aspects are also important for the creation of ashared understanding. Mohammed and Dumville (2001) adopted the term transactive memory fromWegner (1987), which is a set of individual memory systems,which combines the knowledge processed byparticular actorswith a shared awareness aboutwho knowswhat. Transactivememorymakes it possibleto develop complex products with actors from different disciplines without having too muchredundancy of knowledge. With the use of these insights, we dened shared understanding as: asimilarity in the individual perceptions of actors about how the design content is conceptualized (content)or how the transactive memory system works (process).
Earlier research on shared understanding in Co-NPD teams discovered that a lack of a sharedunderstanding caused unnecessary iterative loops in Co-NPD projects (Hill et al., 2001; Valkenburgand Dorst, 1998). Ultimately, a lack of a shared understanding reduces the quality of the nal productbecause, in the end, actors did not solve all problems. Therefore, effective communication is anelementary component of Co-NPD (Bucciarelli, 1996; Hill et al., 2001; Kratzer, 2000; Leenders et al.,2007). Song et al. (2003) added that the highest quality products came from teams with an increase inshared understanding. When there is no shared understanding at the start of a Co-NPD project, actorsneed to discuss the issues at hand and need to learn from each other. Weick and Roberts (1993)explained that groups are smartest in the early stages of Co-NPD because teams loose mind andinterrelating across the disciplines becomesmore routine, more casual andmore automatic at the endof the project, which reduces the quality of the interaction. This could also be an explanation for theremarkable nding of Cummings and Teng (2003) who found that, during Co-NPD projects, non-embedded knowledge transfers were more successful then embedded knowledge transfers.
Previous research showed the importance of the process of creating a shared understandingbetween actors in Co-NPD processes. Research has also shown that creating a shared understanding isdifcult since actors in Co-NPD projects come from different disciplines. However, previous researchdid not show what factors hampered (=barriers) or enabled (=enablers) the process of creating ashared understanding. Yet, these factors are important formanaging knowledge integration processesefciently. This study investigated the barriers and enablers for the creation of a shared understandingin a Co-NPD project in industry. Additionally, we identied relationships between the differentbarriers and enablers and their underlying patterns. We called these patterns collaborativemechanisms. Collaborative mechanisms describe the inuence of the barriers and enablers for thecreation of a shared understanding during Co-NPD projects on knowledge integration.
3. Case description
We studied a Co-NPD team that was responsible for the development of tunnel installations for theDutch high-speed train from Amsterdam to Brussels. The team developed, for example, the elevators,vans and Escape Doors in the tunnels. All tunnel installations need to function together (in the case of
M. Kleinsmann et al. / Journal of Engineering and Technology Management 27 (2010) 203222
an accident), so collaboration was a major factor during this development project. Since there wereeleven technical subsystems, many different disciplines were involved in the team. The Co-NPD teamobservedwas part of a consortium that is responsible for the superstructure of the route and the futuremaintenance.
From a surface level observation of the Co-NPD team, one cannot speak of the development team.The team was a collaborative team hierarchy, from which the structure originated from the newproducts architecture (Gerwin and Moffat, 1997). The team studied consisted of three different typesof sub-teams. The rst type of sub-team was the homogeneous development team that developed atechnical subsystem. The Co-NPDproject studied consisted of eleven of such teams. The second type ofsub-teamwas the system engineering teamwith representatives from the homogeneous developmentteams and two system engineers. This team integrates the eleven subsystems. The third type of sub-team was the multidisciplinary management team, also called the core team. Besides representativesfrom the homogeneous development teams, actors from Quality, V&V (Validation and Verication),RAMS (Reliability, Availability, Maintainability and Safety), Occupational Health and Safety andProcurement were in the core team. The core team planned and monitored all activities needed todevelop the new product. Together, the Co-NPD team consisted of about 60 actors at the time of thisstudy.
4. Research method
The learning history methodwas used as the research method for executing the case study (Kleinerand Roth, 1996; Roth and Kleiner, 2000). This learning historymethod has been developed to facilitateorganizational learning. The method is based on storytelling. Since stories enhance relating events toeach other, they enable the construction of relationships between events. Knowing these relationshipswill enhance the development of an understanding of the mutual relationships between the factorsthat inuence the creation of a shared understanding.
Fig. 1 shows the research steps that were distilled from the learning history method (Kleiner andRoth, 1996). Three basic research steps, data gathering, data processing and data analysis, were used inorder to structure the research process. Each phase consisted of activities and results. In Fig. 1, theactivities are represented by squares and diamonds present the results. The remainder of this sectiondescribes how the different steps were executed in this study.
During data gathering, gaining noticeable results was the rst activity. According to Kleiner andRoth (1996), noticeable results are outcomes, activities, events, behaviors or polic...